Genocide Studies and Prevention: An
International Journal
Volume 2 | 2007
Issue 3 | Article 6
The Order of Genocide : The Dynamics of
Genocide in Rwanda
Scott Straus
Abstract.
To begin, I’d like to describe the origins of the research project that ultimately became The Order of
Genocide. I first traveled to Rwanda as a journalist in the mid-1990s, and, during my travels there
and in what was then Zaire, I became deeply interested in the dynamics that ultimately led to the
1994 genocide. From my perspective as a journalist traveling in the region, the genocide was an
absolutely seminal event, but one that was poorly understood. Eventually I returned to the United
States and began a PhD program in political science at the University of California, Berkeley, where I
began to study the origins of genocide and mass violence, African politics, and the history of the
Great Lakes region. The Order of Genocide is based on my PhD dissertation.
Follow this and additional works at: http://scholarcommons.usf.edu/gsp
Recommended Citation
Straus, Scott (2007) "The Order of Genocide : The Dynamics of Genocide in Rwanda," Genocide Studies and Prevention: An
International Journal: Vol. 2: Iss. 3: Article 6.
Available at: http://scholarcommons.usf.edu/gsp/vol2/iss3/6
This Article is brought to you for free and open access by the Tampa Library at Scholar Commons. It has been accepted for inclusion in Genocide
Studies and Prevention: An International Journal by an authorized administrator of Scholar Commons. For more information, please contact
[email protected].
The Order of Genocide: The Dynamics of
Genocide in Rwanda
Scott Straus
Department of Political Science, University of Wisconsin, Madison
Introduction
To begin, I’d like to describe the origins of the research project that ultimately became
The Order of Genocide. I first traveled to Rwanda as a journalist in the mid-1990s,
and, during my travels there and in what was then Zaire, I became deeply interested in
the dynamics that ultimately led to the 1994 genocide. From my perspective as a
journalist traveling in the region, the genocide was an absolutely seminal event,
but one that was poorly understood. Eventually I returned to the United States and
began a PhD program in political science at the University of California, Berkeley,
where I began to study the origins of genocide and mass violence, African politics,
and the history of the Great Lakes region. The Order of Genocide is based on my
PhD dissertation.
Three related questions drove my initial inquiry. First, what explains the ability of
Rwandan elites to mobilize a large number of citizens effectively and quickly during
the genocide? The literature on African states consistently indicates that governments
are relatively weak, especially outside the capital. Yet in Rwanda state authorities
successfully solicited the participation of hundreds of thousands of citizens in an
extermination campaign. Second, what explains the participation of ordinary civilians
in the genocide? When I began to research Rwanda in the late 1990s there was much
speculation but little by way of satisfactory answers to this question. Yet given the
scale of participation, understanding the dynamics and conditions driving individuals
seemed crucial. Third, why did the elites choose the strategy they did? The literature
on Rwanda was clear that governing, military, and party officials fomented mass
violence; the genocide was not a spontaneous eruption of hatred. Moreover, an
instrumentalist consensus dominated the literature: elites were said to have
deliberately advocated violence to protect and promote their interests. But instrumentalist hypotheses go only so far; they do not say why elites chose a strategy of
genocide over alternatives.
As I considered these questions, a separate set of concerns began to occupy me.
In particular, I became concerned about the gap in detailed evidence about the
genocide. Most existing empirical information was anecdotal or focused on the top
(such as discussions of the history of ethnicity or examinations of national-level
propaganda before and during the genocide). There was relatively little empirical
information about the dynamics of the genocide in rural areas, where the absolute
majority of violence occurred.1 In short, the situation I found as I began to research
Rwanda was that a number of key questions remained unanswered and there was
little systematically collected evidence that could be used to evaluate different
hypotheses. To compensate, I sought to create a research design that would allow me to
generate new empirical information and to evaluate different arguments.
As many who read this journal know, generating evidence about the dynamics of
genocide presents specific problems. Three in particular stood out for me as I started
Scott Straus, ‘‘The Order of Genocide: The Dynamics of Genocide in Rwanda,’’ Genocide Studies
and Prevention 2, 3 (November 2007): 259–264. ! 2007 Genocide Studies and Prevention. doi:
10.3138/gsp.2.3.259
Genocide Studies and Prevention 2:3 November 2007
my research. First, in the aftermath of genocide and the wars in the Congo, trauma
and fear were widespread. Second, narratives were highly politicized, and individuals
across the political spectrum presented highly interested and stylized accounts. Third,
because the topic was the subject of criminal prosecutions, the research questions
had inevitable legal and political implications and were also of concern to institutional
oversight boards, with whose regulations I, like many US-based researchers, had
to comply.
Research Design
Nonetheless, I designed a field research program that consisted of three main stages.
Because my questions focused on the dynamics of violence, I decided to focus primarily,
though not exclusively, on those who took part in the killing. The first and main
research phase was a nationwide survey of imprisoned perpetrators. For the survey,
I used a stratified random sampling method to select and interview sentenced
perpetrators who had pleaded guilty in Rwanda’s domestic courts. The central
research instrument was a semi-structured questionnaire, which I designed
to evaluate competing hypotheses about the dynamics of violence and participation
in it. In total, using this method, I interviewed 210 prisoners in fifteen central prisons
across Rwanda.
The second research phase involved comparing the dynamics of genocidal violence
in five Rwandan communes (the main local unit of administration in 1994). During
my first research phase, I discovered that the patterns of mobilization that led to the
onset of genocide varied from commune to commune (more on this below). In addition,
there was one commune under government control where genocide did not occur.
For this second research phase, therefore, I selected four communes that exhibited
variation in how genocide began as well as the one commune under government
control, Giti, where genocide did not take place. I then studied the dynamics of violence
in each commune in depth through interviews with perpetrators, survivors, current
and former government officials, and non-participating Hutus.
The third stage of research entailed return trips to prisons. During the first two
phases, one clear pattern had emerged: in most communities, a nucleus of core
perpetrators was central to how the violence unfolded. While some of those individuals
fell into my random sample in the first phase, most respondents were lower-level
perpetrators. Thus, for the third research phase, I selected particular individuals to
interview, ranging from top communal authorities to young men who were particularly
active killers in the genocide. I interviewed about nineteen individuals using
this method. In addition to this field research, I also collected as much information
as I could from secondary sources, such as court documents, human-rights reports,
government studies, journalist accounts, and scholarship.
Principal Findings
One important finding from my research is that there was significant regional and
local variation in when and how the violence started. Immediately after President
Juvénal Habyarimana’s assassination—the trigger for the genocide—national hardliners seized control of the Rwandan state and called for the killing of all Tutsis
and prominent Hutu political opponents. However, that call did not translate into
actual violence at the same time in all regions. In some areas, genocidal violence began
within days of the president’s assassination; in other areas, however, it did not begin
260
The Order of Genocide
until four or five days after his death, while in still others the violence took two
weeks or more to get started. In many areas, Hutu leaders and ordinary civilians
initially responded negatively to calls from the central government to kill Tutsis.
By and large, the areas that initially resisted efforts to start the violence were those
where the domestic Hutu political opposition had the most support. Genocidal violence
ultimately began, and succeeded, in those areas, but not before an important delay and
not without a power struggle among Hutus.
In terms of how violence began, significant variation existed among local areas.
In some places, local administration officials clearly started and directed the killing.
In other areas, however, the push came from military officers. In still others, political
party leaders and militia took charge. In some areas, local elites outside the official
administrative hierarchies mobilized to assume control. And in many areas some
combination of these occurred: soldiers worked with local officials or militia and the
rural elite worked with political party officials.
The point to emphasize is that, even though the genocidal outcome was similar
across Rwanda, how and when the violence started varied considerably. Not all Hutus
responded in the same way to the call to commit genocide. That finding is inconsistent
with some common hypotheses about the genocide, particularly the idea that the
decision to take part in the genocide stemmed from a widespread racist culture and
indoctrination; it is also inconsistent with the claim that the genocide was seamlessly
and hierarchically orchestrated, with local officials and peasants blindly following
orders. In fact, the spread of genocidal violence looked more dynamic, like a cascade
of tipping points, rather than being meticulously prepared and implemented.
I collected demographic information from perpetrators, including age, occupation,
number of children, literacy, and years of education. The general finding is that the
perpetrator profile, judging from my sample, was quite similar to that of the adult male
Hutu population in Rwanda at the time of the genocide. In other words, on the whole,
the perpetrator population was not comparatively younger, more unemployed,
or better educated. Rather, the perpetrators were average Rwandan men.
With respect to ethnicity, I found little evidence of widespread interpersonal ethnic
hatred. On a series of indicators—interethnic personal relations, attitudes toward
ethnic intermarriage, and family connections to Tutsis through intermarriage—the
survey yielded consistent evidence among respondents of positive pre-genocide ethnic
interaction. That said, many respondents spoke in categorical terms about ‘‘Hutus’’
and ‘‘Tutsis,’’ in particular when discussing the very tense period after Habyarimana’s
assassination. Some respondents also repeated elements of the genocidal regime’s
propaganda, such as the idea that the rebels killed Hutu children and disemboweled
pregnant women. On the other hand, certain commonly cited propaganda elements
had limited resonance. For example, less than 10% of respondents had heard of—let
alone respected—the infamous ‘‘Hutu Ten Commandments.’’
To investigate the issue of motivation, I asked respondents direct and indirect
questions about how and why they chose to participate in the genocide. As expected,
respondents expressed a range of motivations, from looting to joining attacks to avoid
suspicion that they were hiding Tutsis in their homes. However, two main types of
motivations emerged from these interviews. The first was some form of intra-Hutu
intimidation. Many respondents said that once the violence started in their
communities, they faced strong pressure from other Hutus to participate and feared
negative consequences for themselves and their families if they refused. The second
most common motivation was war-related: respondents said that, in the aftermath of
261
Genocide Studies and Prevention 2:3 November 2007
Habyarimana’s death, they feared that the Rwandan Patriotic Front (RPF) was
invading Rwanda and was killing Hutus.
This discussion of my findings is very brief. Considerably more detail, evaluation,
and triangulation of the findings, as well as historical analysis of prior periods of
violence, can all be found in the book.
The Argument
In The Order of Genocide, I conclude that three main factors drove the genocide
in Rwanda. First, war: I argue that the context of war was critical for the outcome.
War provided the essential rationale for mass killing: security. The logic of Rwanda’s
genocide was predicated on eliminating a threat, on self-protection, and on
re-establishing order. Moreover, the war that took place during the genocide was
intense and defensive, and thus it created a climate of acute uncertainty and
insecurity, sometimes a feeling of panic. That context was critical in motivating some
individuals to foment violence; it was also a key reason that individuals who gravitated
toward extreme measures gained the upper hand wherever they were.
The assassination of Rwanda’s president was part of this dynamic. The
assassination ruptured Rwanda’s political order and thereby created a temporary
gap in authority. The president’s death independently caused anger, leading to calls
for violent revenge; the assassination augmented the anxiety, fear, and confusion of
the war; and the rupture in political order also set the stage for local power struggles.
This last factor is particularly important. After the president’s death, Hutu hardliners
succeeded in gaining control of the state and urged war against the Tutsi ‘‘enemy.’’
That idea—war against the Tutisis—then became the terms around which local actors
asserted power and authority in their communities. The hardliners and those who
adhered to the program of genocide ultimately won the upper hand in almost all areas
not yet lost to the rebels. But such would not likely have happened outside a context
of war, including the president’s assassination.
Given the importance of war and of the president’s assassination, as well as my
findings about how the genocide unfolded at the local level, I conclude that a dynamic
of escalation was a critical factor in the genocide. That, in turn, leads me to argue
that while the genocide was organized, systematic, and ultimately promoted by
the hardliners who took control of the state after the president’s assassination,
the genocide—meaning here the countrywide extermination of the Tutsi population—
was not necessarily meticulously planned in advance, as is often claimed. To be clear,
I argue that the Hutu hardliners who controlled the state are responsible for
the actions they took and, ultimately, for the genocide; however, the dynamics that led
them to foment mass violence, and to succeed in doing so, were in part situational and,
in particular, had to do with the course of the war, including the president’s
assassination.
Second, Rwandan state institutions are critical to understanding why genocide
happened and the participatory character of the violence. The Rwandan state matters
for a number of reasons. First, the state has unusual depth and resonance at the
local level in Rwanda, which meant that, by controlling the state, the hardliners
had the capacity to enforce their decisions countrywide. Second, control of the state
allowed the hardliners to associate killing Tutsis with authority, thus equating
violence with de facto policy. Third, Rwanda has a long history of obligatory labor, and
expectations derived from that history contributed to large-scale civilian mobilization
during the genocide. The potency of the Rwandan state cannot be taken for granted,
262
The Order of Genocide
especially because most African states are weak—particularly in rural areas. Thus, in
addition to demonstrating the importance of the state to the outcome of genocide, I also
explain in the book why Rwanda’s state is so effective at civilian mobilization. Here I
emphasize Rwandan political history, dating to the precolonial period, as well as the
country’s dense geography.
Therein lies a tension. In The Order of Genocide I make the case that insecurity,
uncertainty, anger, and fear related to war and to the president’s assassination drove
the promulgation and spread of violence. However, once coalitions of actors emerged
to win effective control in their communities, they drew on the power of Rwanda’s
local state and the resonance of authority to unleash violence quickly and effectively.
The nature of Rwanda’s institutions and geography also limited exit options, which
both drove high rates of Hutu participation and limited escape opportunities for
Tutsis. The result was a very rapid killing campaign.
Third, ethnicity mattered, but in surprising ways. Overall, I found that ethnic
prejudice, ethnic antipathy, manipulation by racist propaganda, and nationalist
commitments were not the primary drivers. However, the logic of extermination in
Rwanda depended on the idea that Tutsis are of a piece. The genocidal mandate from
the hardliners was to equate ‘‘enemy’’ with ‘‘Tutsi’’ and to declare that Rwanda’s
‘‘enemies’’ had to be eliminated. I argue in the book that the mechanism that allowed
this process to happen is collective ethnic categorization. In case after case, when
justifying killing civilians, perpetrators substituted the category ‘‘Tutsi’’ for the
individuals they were attacking.
The hardliners did not create this category from thin air. Ethnic and racial
categorization has a deep and significant political history, dating, in particular, to
the early colonial period and extending in periodic but pronounced ways into the first
two postcolonial republics. In short, the ethnic/racial categories were ingrained, even if
ethnic hatred was not, and those categories ultimately were essential to the character
of violence—to the fact that violence in Rwanda became genocide.
What caused a shift from an awareness of ethnic categories to collective
categorization and violence? I argue the principal mechanisms had to do with
uncertainty, fear, social pressure, and opportunity. In the aggregate, Hutus
participated in genocide because they wanted to protect themselves during a war
and a period of intense uncertainty; because they felt that complying with those who
told them to kill would be less costly than not complying; and because they
opportunistically used the period of confusion and violence to obtain power and
property. These dynamics, I argue, are inseparable from the specific context of war and
from the nature of Rwandan state institutions and geography.
Implications for Post-Genocide Rwanda
Like the presentation of my findings, my argument here is truncated.
In closing, I would like to discuss briefly—again, more detail is in the book—the
implications of my findings and argument for post-genocide Rwanda.
Rwanda’s RPF-dominated post-genocide government has favored strong control
over public political discourse and maintained a strong military emphasis. These
policies are based, at least in part, on an interpretation of mass participation in the
genocide that tends to see the genocide as an undifferentiated event and
the main cause of participation as mass beliefs and mass racist indoctrination.
This interpretation essentially posits a persistent post-genocide threat, because the
Hutu population is considered to be either genocidal or brainwashed and thus prone
263
Genocide Studies and Prevention 2:3 November 2007
to becoming genocidal. The logical response to such an ongoing threat is strong
coercive control over the potentially dangerous population.
However, there is substantial risk that a strong security outlook, even though it
will keep the peace in the short term, will alienate large portions of the population and
sow the seeds of instability in the long term. Here the evidence I collected could
contribute to rethinking policy. First, the evidence does not support collective blame of
the Hutu population. Not all Hutus were ge´nocidaires, and not all perpetrators
participated to the same degree. Second, the genocide was not perpetrated in a
seamless, ‘‘machine’’-like manner. Rather, the violence was the outcome of local-level
struggles for dominance, and many initially resisted or tried to avoid becoming
involved in it. Third, the evidence suggests that while ethnic categories were an
important background condition for the genocide, neither interpersonal ethnic enmity
nor a deeply imbibed racist culture was the wellspring of most individuals’
participation in the violence. Rather, specific conditions in Rwanda triggered the
saliency of ethnic categories and enabled a dynamic of violence to take hold. Thus, in
considerations of how prone Rwanda is to future violence, the stress should fall on
those situational factors that sowed the seeds of insecurity and destabilization that
allowed a dynamic of violence to take root.
All this indicates that some key conditions that facilitated the genocide in the first
place are no longer present. By extension, the prospects for post-genocide confidence
among social groups may be greater than many Rwandan and outside observers fear.
Nothing about social reconstruction after mass violence is easy, but my evidence shows
that Rwanda is not, and was not, a nation of people predisposed to violence.
Acknowledgments
I would like to express my sincere thanks to Samuel Totten for publishing this book
forum and to the three commentators.
Note
1.
An exception was Alison Des Forges’ impressive human-rights documentation: Alison Des
Forges, Leave None to Tell the Story: Genocide in Rwanda (New York: Human Rights
Watch, 1999).
264